When the second argument contains batch dimension, the axis should be 0.

Also makes elementwise ops more tolerant at handling tensors with trailing
singular dimensions.

fetch_var for getting the values of a variable with given name
get_var for getting the Variable with given name

* remove pick dependency

* fix bug

* small fix

* modify executor.py for save and load

* clean code

* Add usage example

* refine executor run function

* fix bug

* refine executor code

* fix block bug

* fix comments

* fix bug

* fix pass num

* "fix decode bug"

* "follow commnet"

* "fix error"

* "fix hook bug"

* fix based comment

* fix copyright

* fix based on comment

* add copyright hook

* add copyright hook

* refine copyright hook

* "test copyright hook"

* fix check style

* fix ci

* "remove cudnndevicecontext"

* "remove unused init code"

* "fix hash functions"

* "change operator interface"

* "move devicepool to device_context"

* "fix operator test"

* "fix op_registry Run interface"

* "net op passed. Need to fix nccl multi-Context"

* "add nccl group function"

* "add nccl group function"

* "fix gpu count exceed 32 error"

* "fix recurrent op, nccl op"

* "change the other operators interface with Place"

* "fix typo"

* "fix pybind"

* "fix device in python side"

* "fix pybind failed"

* "add init for test"

* "fix CI"

* "add DeviceContextPool"

* "add devicecontextpool in pybind"

* "add comments in python side "

* "fix static link error"

* "fix CI error"

* "add executor.py"

* "fix CI error"

* "add with gpu macro"

* "remove comment out codes"

* "add TODO items"

* "update init devices"

* Unify fluid submodules to fluid module

Change books just use `import fluid`, not submodules

* Remove g_main_program/g_startup_program

Use default_main_program/default_startup_program instead

* Typo

* Fix CI

Change books just use `import fluid`, not submodules

* "add asnumpy interface"

* Just for unittest

* Change unittests for numpy I/O

* Fix CI

* init commit

* change some dir name

* Use stable_sort in lod_rank_table

It is easy to debug and test when use `stable_sort`and the time
complexity is not changed.

* Add LoDTensorArray

* Stash

* Better debug message for IsInitialized

* Stash

* Better debug message for IsInitialized

* Complete array read/write op unittests

* Init commit

* Make executor use ProgramDescBind

* Change Attribute from BlockDesc to BlockDescBind

* Since we will get the program desc in RNN, just BlockDesc is not
  enough.

* Add DeviceContext to Executor API

* Rewrite RNN

* Pass Python

* AddBiasOp does not care num_flatten_dims

* Stash

* Fix MacOS Compile

* Pass RNN forward

* add python test

* refactor test

* Make compile pass

* add gradopmaker

* First draft done

* Polish code

* add grad op maker and grad infershape

* Polish code

* Fix backward.cc bug

* Fix infershape

* Rename function

* add backward test

* simplify recurrent test

* Update

* Pass unittest

* Add comments & refine test

* Add comments

* refactor test

* Complete Unittest

* fix StepScopes enforce

* Remove unused unittest

* no type error

* Update

* Make RNN Pass unittest

* Python API for save/load variables

* Polish names

* Unify `set_feed_variable` to one method

* Move global scope to python, not in C++